Coding Agents

This section describes how to use coding agents in Asterinas contribution workflows. Agents can accelerate implementation, investigation, testing, and review, but they must operate inside a disciplined engineering process.

The goal is not to maximize agent autonomy. The goal is to make delivery more reliable: clearer task boundaries, better intermediate artifacts, stronger verification, and lower review cost.

The material is organized into the following pages:

• Operating Principles — the engineering model for using agents in Asterinas.
• Skills — guidance for packaging reusable agent workflows into stable skills.
• Task Fit — what coding agents are good at, where they are weak, and where trust should be reduced.